Introduction/Background Mismatch repair deficiency (MMRd) accounts for 20–30% of all endometrial cancers (EC). MMRd ECs are characterized by the loss of at least one of the MMR proteins (MLH1/PMS2/MSH2/MSH6) at immunohistochemistry (IHC), with MLH1-PMS2 and MSH2-MSH6 loss occurring mostly as heterodimers. The MMRd group has been associated with a good-intermediate prognosis. However, the value of the different MMR loss patterns has not been explored. We aim at characterizing how MMRd protein expression patterns are associated with clinicopathological data.

Methodology Clinicopathological data from surgically staged EC patients have been retrospectively collected, with MMR status/p53 mutational status/estrogen receptors status (ER)/progesterone receptors status (PR) being evaluated by IHC. The population was divided into three cohorts: loss of MLH1, PMS2 or both (MLH1/PMS2 group); loss of MSH2, MSH6 or both (MSH2/MSH6 group); positive IHC staining for MLH1, PMS2, MSH2 and MSH6 (MMR-proficient group, MMRp).

Results Data from 1451 EC patients were analyzed. We identified 1075 cases (74.1%) with MMRp EC, and 376 (25.9%) with MMRd EC. Among the latter group, we observed 314 (21.6%) patients with MLH1, or PMS2, or MLH1 and PMS2 negative (MLH1/PMS2 group), and 62 (4.3%) patients negative for MSH2, or MSH6, or for both proteins (MSH2/MSH6 group). The MLH1/PMS2 cohort was older (p<0.001), with a higher BMI (p<0.001) and a trend toward a higher presence of nodal involvement (p=0.05) compared to MSH2/MSH6 patients. MSH2/MSH6 patients displayed a higher concurrent p53 mutational rate (p=0.016). When compared with the group of MMRp, MLH1/PMS2 had an increased rate of FIGO stage IIIC1 and IIIC2 (p=0.001), nodal involvement (p<0.001), lymphovascular space invasion (p=0.007), deeper myometrial invasion (p=0.002), and larger tumor dimensions (p=0.002). Comparison for the same pathological features between the MSH2/MSH6 and MMRp groups did not detect any difference.

Conclusion The MMRd MLH1/PMS2 pattern is associated to higher risk clinicopathological features compared to the MHS2/MSH6 counterpart.

Disclosures None